2.0 Analysis 2.1 Class Certification Status From the completion of the second routine annual survey on 27 November 1997 until the time of the occurrence, the FLARE operated with an Interim Certificate of Class, issued subject to a COC that required certain structural repairs to be completed before the end of February 1998. The imposition of a COC calling for the completion of remedial action at the next routinely scheduled survey, or within a shorter specified time, is a long-established and universally adopted practice of classification societies. The duration of any such deferment of repairs is allotted by class surveyors and is based, in part, on the perceived importance of the particular deficiency and the availability of suitable repair facilities at the current location. Depending on the nature of the deficiency, class surveyors may require that remedial action be carried out immediately or, in the absence of suitable local repair facilities, at the nearest, or the vessel's next suitably equipped, port of call. Class surveyors may also, with propriety, impose operational limitations on the vessel until the remedial repairs are satisfactorily completed. None of these actions was deemed necessary in this instance. It is the responsibility of the owners to ensure that all repairs necessary for the maintenance of class are carried out and that the classification society is advised of the proposed scheduled venue, in order that a surveyor may attend for inspection and approval purposes. At the time of the occurrence, no formal request for the provision of such services had been received by the Lloyd's Register of Shipping. The FLARE discharged cargo in the fully equipped port of Rotterdam prior to departing in a lightly ballasted condition on 30 December 1997. The vessel subsequently broke in two and sank while crossing the North Atlantic Ocean in typical winter weather--still within the three months allotted by the COC and before remedial repairs to the internal structure of several upper wing water ballast tanks were completed. 2.2 Port State Control PSC inspections in the USA and Canada did not result in the vessel being detained, whereas the inspection in Newport, Wales, did. The crew members were usually changed on a rotational basis and those crew members who performed poorly at the first lifeboat drill in Toronto were not necessarily the same as those on board at an earlier or later date. The refusal of TC officials to accept the standard of the crew's performance at the boat and fire drill held on 26 May 1997 indicates that the crew (at that time) was not practised in the use of this equipment. This may be an indication of inadequate levels of shore-based pre-sea training and on-board instruction. The PSC inspection in Newport found, inter alia, that some lifeboat and emergency equipment was deficient, crew accommodation was used for the storage of inflammables, and the accommodation was infested and unsanitary. These findings point to insufficient maintenance, dangerous practices, and poor housekeeping. 2.3 ISM Code Certification ISM Code accreditation of the vessel and the company had not been completed at the time of the occurrence. Preparations were well advanced toward the formulation of a company Safety Management System prior to an audit to obtain a Document of Compliance and a Safety Management Certificate for the FLARE. Although the vessel's classification and regulatory documentation was in order, it did not address the application of existing operational procedures. A Safety Management System would have referred to the appropriate ballasting procedures contained in the vessel's Loading Manual. These ballasting conditions would have reduced the vessel's vulnerability to pounding--given that the vessel was about to embark upon a crossing of the North Atlantic in winter. 2.4 Pilots' Observations Although the information gleaned from the harbour and river pilots is of an anecdotal nature, none of their observations gives rise for concern regarding the seaworthiness of the vessel or her handling characteristics. 2.5 Vessel Ballasting and Trim The light ballast loading condition and draughts recorded in Rotterdam prior to departure, and the subsequent comments of the port pilot on departure, confirm that the vessel sailed with a marked after trim and a relatively shallow forward draught. The forward draught and the total quantity of water ballast on board were lower than those shown for the light ballast departure condition in the Loading Manual, which was provided for the guidance of the master. Because of the various reported ballasting changes during the voyage (due to ongoing repairs in way of the upper wing tanks), the precise longitudinal and transverse distribution of the water ballast immediately prior to the hull failure is not known. However, the forward and after draughts of 11 feet (3.35 m) and 21 feet (6.4 m) reported by the master to ECAREG on 13 January 1998 clearly indicate that no substantial amount of additional water ballast had been taken on board at that time, that the afterpeak and the deep tank remained empty and the forepeak tank was less than full. Furthermore, the slight reduction of after trim indicated by the last reported forward and after draughts is consistent with the consumption of fuel oil and other liquids, etc., from their respective tanks, all of which were aft of midships. Canadian guidelines regarding exchange of ballast water, for ships proceeding to the St. Lawrence River and the Great Lakes, recognize that such exchange may be impracticable in exceptional circumstances (such as extreme weather conditions). The exchange may then be deferred until the vessel enters more sheltered waters. In this instance, the master reported that seawater ballast had been exchanged on January 10, 11 and 12, this being the only period of less severe weather experienced on this passage. In his response to the vessel's operators, the master further reported that the ballast tanks were full. However, an analysis of photographs and reported draughts shows that the tanks were less than full. A common factor throughout the recorded and reported loading conditions is that the forward draught throughout the voyage was consistently shallower than any of the ballasted departure conditions given in the vessel's Loading Manual. The forward draught was also less than that contained in the Lloyd's Rules and Regulations for the Construction and Classification of Steel Ships. The minimum forward draught so indicated has been found satisfactory--and proven by extensive operational experience--for the prevention or reduction of hull pounding. The vulnerability of the vessel to pounding and slamming (and the repetition of such impacts over several days, as reported by the survivors) was a direct consequence of the vessel's shallow forward draught. The resulting shock loads imparted to the hull girder were the major factor in creating sudden localized stress levels greater than the structure could withstand, and led to the sudden catastrophic structural failure. It is the responsibility of the master to ensure that his vessel is safely operated, loaded, appropriately ballasted and trimmed throughout the whole voyage. Details of the deep ballast loading condition were included in the Loading Manual for the guidance of the master; if adopted, they would have markedly reduced the vessel's vulnerability to pounding and slamming. However, it is not known why the deeper ballasting arrangement was not employed. 2.6 Damage and Hull Separation Sequence The vessel's speed, lightly ballasted condition, shallow forward draught and the prevailing rough and stormy weather conditions were such that the forward end of the hull was subjected to repeated pounding and slamming for several days before the final structural failure of the hull. The severity of a very heavy slam incurred by the forefoot and bottom shell plating, some four hours before the hull separation, caused a loud bang and severe longitudinal vibration and whipping of the hull. The vessel apparently continued at the same speed, and it is not known if an inspection to determine possible hull damage was made at that time. Survivors reported that, shortly before the catastrophic structural failure, the vessel experienced a further particularly heavy pounding shock, which was also followed by very severe longitudinal whipping and flexing of the hull. The occurrence and severity of these pounding shocks is most likely due to the vessel having suddenly encountered particularly high or irregular waves. Another vessel, in the same general area at about that time, reported having experienced similar irregular phenomena. The high amplitude and frequency of the hull vibrations resulting from such severe slamming subjected the vessel to the simultaneous effects of great strain energy input and very high stress rate. These induced loadings would cause sudden, rapidly repeated, high stress concentrations in the main deck and bottom shell plating, which form the principal outer members of the hull girder. Such shock loads can create stresses in excess of the maximum approved static levels. Any localized structural discontinuity or existing undetected fissure damage in the outer members of the principal hull girder can further concentrate such loading. This shock loading can result in the rapid propagation of extensive brittle fractures and lead to sudden catastrophic structural failure. Based on the reported loading, trim, prevailing severe weather conditions, hull surveys and review of the aerial and underwater photographic records, it appears most likely that the structural failure sequence was initiated by hull slamming and vibration-induced brittle fractures in the main deck plating. The initial brittle fracturing occurred abreast and adjacent to a grain-loading port on the starboard side of the main deck near midships. The suddenly imposed, concentrated and rapidly repeated stress loading exceeded that which the local structure could withstand, and the resulting fracture rapidly propagated across the deck and continued into the starboard side shell plating. Existing fissure damage found on the inner surface of the radiused gunwale on the port side of the main deck near Frame 76 (see Photo 5) contributed to the rapid propagation of a brittle fracture across the main deck, through the sheer strake, and into the port side shell plating. The reduction of longitudinal structural integrity following the propagation of brittle fractures in the upper members of the hull girder resulted in high compressive loads being suddenly imposed on the bottom structure. Subsequent very high stress level concentrations in way of existing fissure damage in some bottom shell plating then led to the sudden failure of the entire bottom structure and resulted in the catastrophic hull separation. The position of a brittle fracture located in the port side main deck plating near Frame 110 generally coincides with the starboard side main deck failure at Frame 112, and both are virtually at the mid-length of the vessel. Principal structural members near midships are subject to repeated high static and sudden shock-induced stress levels and, consequently, are most prone to structural fatigue as a result of such cyclical loading. In this instance, however, close-up internal access at these locations was precluded, and the presence of existing fatigue fissures in nway of the main deck plating failures nearest to midships could not be verified. The small temperature gradient due to the ambient air and seawater temperatures (-3C and 2C, respectively) was such that any adverse effects from the generation of thermal stresses within the structural members of the hull may be regarded as having been insignificant. Factors, other than those resulting from severe pounding, that contributed to the global and localized increases in stress levels imposed on the hull and main deck plating included: the reduction of load-carrying effectiveness of the deck longitudinals due to the lost or reduced support from the heavily corroded transverse supporting webs inside the upper wing tanks; wracking stresses from the asymmetrical transverse distribution of water ballast; and fluctuations in the longitudinal and torsional hull stresses resulting from the vessel's motion in the stormy sea conditions. 2.7 Multinational Crewing The multinational crewing of vessels is a long-established practice. Where the English language is the common and working language of the ship, problems may arise when non-English-speaking crew do not fully understand instructions or their intent. Such language differences can lead to uncertainty, misunderstanding and a lack of control where circumstances demand immediate action. That this problem existed on board the FLARE was evident. A Yugoslav member of the crew reported that, at emergency drills, the Yugoslav chief engineer had to translate the Greek master's English instructions for him. 2.8 Search and Rescue (SAR)--Background Canada, as a participant in the IMO, SOLAS, the United Nations Convention on the Law of the Sea, the International Convention on Maritime Search and Rescue, and various other Conventions, is responsible for all marine SAR services within its designated, assigned area. Both the Department of National Defence (DND) and the CCG have SAR responsibilities under the National Search and Rescue Program. Within Canada's area of responsibility, all marine SAR services are coordinated through Rescue Coordination Centres (RCC) and Marine Rescue Sub-centres (MRSC), which are strategically located across Canada. The relevant locations in this occurrence were RCC Halifax, Nova Scotia, and MRSC St. John's, Newfoundland. In this occurrence, the marine coordinator of the MRSC, who was appointed SMC, worked closely with RCC Halifax. The air coordinator at RCC Halifax continued to control and coordinate air resources. The Halifax Search and Rescue Region covers an area of approximately 6.1 million km2. It includes the Atlantic provinces to the Canada/USA border and the eastern portion of Quebec, and it extends as far north as Baffin Island and approximately 1,000 miles (1,600 km) into the Atlantic Ocean. In order to ensure timely assistance and the saving of lives through continuous monitoring of international frequencies, Stephenville MCTS has peripheral sites at several locations, including Ramea, Newfoundland. Services include distress and safety communications and coordination to detect distress situations. Stephenville MCTS is not equipped with a VHF/direction finder (DF) to give a line of bearing on Safety, Urgency, and Distress Traffic. On the east coast of Canada, weather conditions that affect the delivery of SAR service include severe sea states and gale force winds, freezing spray, ice cover and fog. During winter storms, wave heights of 30 m and wind speeds of 160 km/h are not unknown. A study conducted in 1992, SAR Needs Analysis,[5] identified the need for two Arun Class lifeboats as primary SAR vessels for area 034 (south coast of Newfoundland). While these vessels are proven performers in coastal waters, they are less suitable when required to operate further offshore, as their limited endurance restricts their offshore SAR patrol capabilities. At the time of the sinking of the FLARE, one of these lifeboats was on station at Burin, Newfoundland, and the other at Burgeo, Newfoundland. Their tasking was delayed until 1200 due to the severe weather and the fact that the lifeboats would be operating at maximum range in the area of the occurrence. The secondary resources nearest to area 034 were the CCGS ANN HARVEY in the Gulf of St. Lawrence and the CCGS J.E. BERNIER. Both vessels were about 240 miles distant and were, respectively, some 15 and 19 hours' steaming time from the scene. 2.9 MAYDAY Reception Although it was determined that the distress alert originated in Canadian waters, due to the indistinct and incomplete MAYDAY transmission, the lack of VHF/DF equipment at the receiving stations and the fact that no EPIRB signal was received, the actual position of the sinking was not determined for some time. The Standards Manual (TP989) in use at Stephenville MCTS on the morning of the occurrence states in section 10.3.1: Requests for information by RCC/MRSC from recorded tapes during or shortly after amarine incident shall not be provided unless the station supervisor or his delegate is available to obtain the required information. In all cases an approved playback unit must be used to obtain the information. The tape recorder was located in a room adjacent to the operations room, and the operator had to leave his/her station to access the recorder. Because of the physical layout of the MCTS station and the operational procedures in place, the lone watch officer was required to call the supervisor back to duty in order to have the MAYDAY tape replayed. Consequently, a delay of 55 minutes occurred in deciphering the latitude transmitted in the indistinct MAYDAY. In the event, the deciphered latitude differed from the latitude in which the bow section was eventually found by 8.3 minutes (nautical miles). The FLARE was not equipped with a global positioning system for navigation. She was equipped with a satellite navigation system, which was not integrated with the Global Maritime Distress and Safety System (GMDSS). The FLARE was not equipped with DSC; under existing regulations, she was not required to be so equipped until 01 February 1999. Had she been so equipped, a distress alert could have been sent by VHF DSC and MF DSC by the press of a button. Instead, the distress call was made by VHF radiotelephone. When the VHF MAYDAY and the INMARSAT C positions were plotted, there was a large discrepancy between them. To obtain a better estimated distress position, MRSC/RCC plotted the last ECAREG position given by the vessel at noon on January 15 and extended it through the 0221 INMARSAT C position (realizing that the INMARSAT C position was suspect at this time). This track line, when extended, was in keeping with the vessel's intended track to Cape Ray. The position where this track cut the latitude arrived at by the supervisor at Stephenville MCTS was then taken as the estimated distress position (4637.15' N, 05800' W) and was therefore used in conjunction with the assumed MAYDAY position, the INMARSAT C position and the position provided by the Canadian Navy MOC. This resulted in a very large search area. Since some uncertainty still prevailed regarding the original assumed MAYDAY position, resources were also sent there, even though the STOLT ASPIRATION, tasked within 10 minutes of the MAYDAY, had reported no targets in the vicinity, making the MAYDAY position even more questionable. Uncertainty led to SAR resources being directed to several different locations over a wide area. Failure to investigate the unidentified radar target when it was first reported by R01 delayed resolution of the uncertainty. R306 did not report the target to RCC and assumed it to be a rescue vessel. The aircraft crew was also preoccupied with reconfiguring data buoys, preparing for a search pattern and handling communications. The misidentification of the radar echo of the STOLT ASPIRATION for that of the FLARE resulted in delay of the determination of a more accurate position of the sinking. The SAR controller recognized almost immediately that, while the 0221 INMARSAT C position was correct, the earth station printout data was incorrect due to a computer software error at the Southbury, Connecticut, earth station. This error was subsequently corrected. 2.10 Lifejacket Testing As only one lifejacket (recovered from a survivor of the FLARE) was available for testing, its condition may not be typical of the vessel's other lifejackets. The lifejacket was tested by the Underwriters Laboratories of Canada. Visual examination showed that some of the tapes had been cut, but the fabric and seams appeared sound. A cord attached to the lifejacket may have been the attachment for a whistle but there was no indication that a light had been fitted as required by the standards. However, during the PSC inspection in Newport, Wales, in October 1997, the vessel's lifejackets were found to be in compliance with SOLAS but for the fact that the dates on the lifejacket light batteries were unidentifiable. A 1.88 m tall male with a body mass of 94.5 kg tested the lifejacket. In donning the lifejacket, some instruction concerning the collar attachment was required. The wearer found that it was uncomfortable and that the collar was restrictive. Before he jumped into the water from a height of 5 m, the wearer elected not to tie the lifejacket collar securing tapes, which would have led across his throat. The jump into the water resulted in the lifejacket riding up his body, with the upper body tapes rising above his shoulders. When in the water, the wearer adjusted the collar securing tapes and he was supported in the required vertical position. The wearer was able to swim and was turned by the lifejacket from face down to face up within five seconds. Buoyancy was not reduced by more than five per cent following 24 hours' immersion. 2.11 Immersion Suits The number of immersion suits required on board cargo ships depends on the number of lifeboats carried. SOLAS regulations require that three immersion suits be provided per lifeboat and sufficient thermal protective aids be provided for the remainder of the crew. If the Administration considers it necessary and practicable, one immersion suit for every person on board may be carried. Further, if the vessel is constantly engaged on voyages in warm climates where, in the opinion of the Administration, immersion suits are unnecessary, none need be carried. Immersion suits that meet the requirements of SOLAS are invaluable when abandoning ship in cold climates. Because of this, all Canadian-registered vessels of the type and size of the FLARE are required to carry an immersion suit, with a whistle and locator light attached, for each member of the crew.[6] Such a requirement for ships operating at any time in cold waters could be a positive step toward saving lives--if crews are familiar with both the location and use of the immersion suits. 2.12 SAR Airborne Response Due to design limitations in terms of power, weight-carrying restrictions are imposed on Labrador helicopters. Certain life-saving equipment is no longer routinely carried unless a specific need for it is identified at the start of a SAR mission. Shortly after Labrador helicopter R304 was tasked and airborne, the crew detected a suspected hydraulic leak in the after transmission and the helicopter returned to Sydney, Nova Scotia, for inspection. It was found that the suspected hydraulic leak was in fact a mixture of melting snow, freezing rain and some residual hydraulic fluid in the after upper pylon and that no leak existed. The participation of the helicopter in the SAR operation was delayed as a result of this precautionary landing. 3.0 Conclusions 3.1 Findings as to Causes and Contributing Factors While the vessel was making a winter crossing of the North Atlantic, the No. 4 hold/deep tank was not filled with ballast as indicated in the deep ballast loading condition of the vessel's Loading Manual. The light ballast loading condition on departure and the shallow forward draught made the vessel highly vulnerable to pounding and slamming. The required structural repairs were not effected before leaving Rotterdam, a port with extensive ship-repair facilities. Throughout most of the voyage from Rotterdam toward Montreal, the vessel encountered westerly gale- to storm-force winds and very high seas. Fractures in the boundaries of the upper wing ballast tanks were discovered and repaired during the voyage, but the internal structural repairs required by the Condition of Class had not been completed at the time of the occurrence. The contents of some of the upper wing ballast tanks were adjusted while riding repairs were being effected, but the precise distribution of water ballast at the time of the hull separation is not known. At about 0000 (ship's time) on 16 January 1998, the FLARE encountered large and steep irregular waves, which were also encountered by another vessel in the vicinity. These waves reportedly caused slamming of the vessel's forefoot, followed by a loud bang and severe hull whipping and vibration. This was followed at about 0430 (ship's time) by another particularly loud bang, again with severe hull whipping and vibration. At about 0430 (ship's time), loss of longitudinal structural integrity was initiated by rapid brittle fractures that occurred in the main deck plating in way of grain-loading ports and existing fissure damage near midships. Bottom structural failure, resulting from suddenly imposed compressive loading and excessive localized stress concentrations in way of existing fissure damage, caused the hull to break in two. 3.2 Findings Related to Risks to the Vessel, to Persons and to the Environment At the time of the occurrence, the vessel was operating with an Interim Certificate of Class because a Condition of Class had been imposed that required structural repairs in the upper wing ballast tanks to be completed before the end of February 1998. It was the master's first voyage in command of a vessel of this size and type, but he had previous experience as first mate on similar vessels and as master of other, smaller vessels. The master and 11 of the crew members had joined the vessel in Rotterdam and had had limited opportunity to familiarize themselves with the ship or her equipment. The crew was comprised of four nationalities; English was the common language of communication, but a survivor reported that he was unable to understand safety instructions without translation. At 0832 UTC (0432 ship's time) on 16 January 1998, Marine Communications and Traffic Services (MCTS) at Stephenville, Newfoundland, received a hurried, indistinct and incomplete MAYDAY from an unidentified vessel on VHF channel 16. Since the MCTS watch officer was not authorized to replay the tape recording of the MAYDAY message, the MCTS supervisor was recalled to the station to assist. There was a delay of some 55 minutes before the incomplete information could be analysed. After the vessel broke in two, the stern section listed 30 to 35 degrees to starboard, precluding the launch of the starboard lifeboat (which broke free some time after the stern sank). The crew was unable to clear away and launch the port lifeboat due to difficulties encountered in freeing extra lashings made to secure the boat in the heavy weather experienced during the crossing. This lifeboat broke or drifted free and capsized when the stern section sank. Great difficulty was experienced in launching a liferaft over the stern due to ice- and snow-covered decks, but the crew did not immediately abandon ship to the liferaft because the vessel's propeller was still turning. The liferaft's painter reportedly chafed through and the liferaft drifted away from the vessel's stern. The liferaft on the foredeck apparently remained aboard. As the stern section sank, the crew, wearing lifejackets, abandoned ship into the sea. Of six who swam to and climbed onto the capsized port side lifeboat, four survived to be rescued. The vessel's Emergency Position Indicating Radio Beacon (EPIRB), inspected and certified satisfactory in Cuba in November 1997, either did not float free or did not self-activate as designed. No signal was ever received from it nor was it recovered. The FLARE was equipped with two search and rescue transponders (SARTs), which were stowed on the bridge deck; however, in the event, no SART response was received by search and rescue (SAR) equipment. Due to continued uncertainty of the position from which the MAYDAY call originated, SAR resources were, at first, tasked to a large area. A position of the FLARE (0221, January 16), obtained from INMARSAT C data, appeared not to have been updated for 24 hours. However, the position was later considered valid, and the confusion arose due to a computer software error at the Southbury, Connecticut, earth station, which was subsequently corrected. Further positional information from all sources was correlated and the area of the search was redefined at 1157. The lifejacket that was recovered from a survivor and later tested did not meet the lifejacket standards of the 1992 Consolidated International Convention for the Safety of Life at Sea (SOLAS), nor was it required for a ship of this age. It is not known if the lifejacket was representative of the condition of the other lifejackets aboard the FLARE. There were six immersion suits on board, but the survivors were unsure of their stowage location and, in the event, the suits were not used. An oil slick caused by fuel oil escaping from the sunken stern section became widespread and non-recoverable. Contamination from the sunken ship's fuel oil on the helicopter work area caused the helicopter's flight crew and SAR technicians (SAR TECHs) to become nauseous, making working conditions unacceptably hazardous. Similar conditions were experienced on the working decks of the surface rescue vessels. 3.3 Findings of an Informational Nature At the time of the occurrence, the certificates of competency of the master and officers, and the qualifications of the crew conformed with the regulatory requirements for this class of vessel, and were appropriate to the service in which she was engaged. At a Port State Control (PSC) inspection in May 1997, in Toronto, the crew at that time, on its first attempt, did not demonstrate its competence in boat and fire drill to the satisfaction of the Transport Canada inspector. At the time of the vessel's last PSC inspection in October 1997, in Newport, Wales, the vessel was detained, but later sailed after deficiencies in life-saving appliances were rectified. The vessel's registry, classification and safety equipment documentation was valid until November 2000. The stern section of the vessel sank about half an hour after the hull separated, in a position some 45 nautical miles southwest of the islands of Saint-Pierre-et-Miquelon; the bow section remained afloat for four days, and sank some 80 nautical miles southeast of Louisbourg, Nova Scotia, on 21 January 1998. Four survivors were sighted clinging to the capsized lifeboat at 1423, and all had been rescued by a SAR helicopter by 1434. The extensive airborne SAR response involved a chartered commercial fixed-wing aircraft equipped for aerial surveillance; five Department of National Defence (DND) SAR fixed-wing aircraft; and four DND SAR helicopters. The seaborne SAR response involved two commercial vessels; five Canadian Coast Guard (CCG) vessels; one Canadian Navy ship; and one French patrol vessel. At the conclusion of the search, of the crew of 25 men, 6 remained missing, 15 bodies were recovered and there were 4 survivors. 4.0 Safety Action 4.1 Action Taken 4.1.1 Warning Against Possible Similar Defects Preliminary analysis using the aerial photographs and the underwater survey of the vessel indicated that fissure damage and failure in the main deck plating adjacent to grain-loading ports near the mid-length of the FLARE may have existed prior to the hull failure. The TSB identified 14 vessels of similar age and built to the same plans that were believed to be still in service and that could be subject to similar defects. Consequently, in September 1998, the TSB apprised all involved Flag States of the fissure damage on the FLARE, in order that such defects on similar vessels under their administration might be identified in a timely manner, and appropriate remedial action taken. The vessels identified were: * Those vessels marked by an asterisk were not classed with the Lloyd's Register of Shipping and consequently were not inspected by that classification society. Flag States' feedback was important in considering findings as to causes and contributing factors leading to this occurrence. Responses from the Flag States indicated that noted vessels in operation under their jurisdictions had been either inspected or caused to have been inspected by the appropriate classification societies and were found in satisfactory condition. (One vessel had been scrapped in July 1998.) The Lloyd's Register of Shipping also required Unscheduled Surveys on sister ships, classed with the Lloyd's, to determine possible structural defects. The TSB also communicated safety information to the international media on the occurrence. For example, The Motor Ship magazine provided coverage on the occurrence and safety issues to the global maritime community in its November 1998 issue. 4.1.2 Action Taken by Transport Canada - Marine Safety (TCMS) The TSB also apprised Transport Canada of its concerns, in order that its Port State Control (PSC) inspectors could take any necessary action during their inspection of similar vessels or the vessels listed above. Two of the vessels arrived in Canada and were inspected by TCMS. Both were detained; one with structural defects similar to those of the FLARE and the other for defective life-saving equipment, navigation equipment and tank remote shut-offs. Both were released after these deficiencies were corrected. In addition, TCMS has forwarded the list of similar vessels to the Marine Communications and Traffic Services (MCTS) to be included in the Ships of Particular Interest (SPI) List in order that these vessels, when reported inbound, may be placed under special surveillance under the PSC inspection program. In 1998, TCMS also participated in the Paris MOU Concentrated Inspection Campaign on Bulk Carriers. The campaign was aimed at the inspection of the structural safety of bulk carriers which are more than 30,000 gross tons and more than 15 years old, particularly those carrying high-density or corrosive cargoes and trading on the spot market. Subsequently, TCMS issued two Ship Safety Bulletins (SSB); No. 20/98 Sea Coasts of Canada - Global Maritime Distress and Safety System (GMDSS) to inform mariners on GMDSS, and SSB No. 13/98 Code of Practice for the Safe Loading and Unloading of Bulk Carriers. The Code was developed by the International Maritime Organization (IMO) to assist persons responsible for the loading or unloading to carry out their function safely and to promote the safety of bulk carriers. 4.1.3 Search and Rescue (SAR) Operation's Report and Recommendations Following the completion of SAR operations, the Search and Rescue Mission Coordinator (SMC) for the FLARE operations (St. John's Marine Rescue Sub-Centre) produced a comprehensive report on the conduct of the operation. Based on the lessons learned, the SMC made five recommendations to further improve SAR operations in the future. The report recommends in part: That Atmospheric Environment Services implement a high definition analysed forecast wind field to improve the accuracy of the search area calculated by the Computerized Search and Rescue Planning Tool (CANSARP). That all Canadian Coast Guard (CCG) MCTS very high frequency (VHF) sites be fitted with VHF/direction finder (DF) to give a line of bearing on all VHF Safety, Urgency and Distress Traffic; That an offshore CCG Marine Resource for SAR purposes be stationed off the Newfoundland South Coast and a cooperative agreement be put in place to utilize existing resources stationed in the area; That all MCTS radio receiving sites immediately install a digital callback machine so as to gain immediate playback capability of all distress traffic; and That the Search and Rescue Communications (SARCOMM) Land Lines be reactivated at all MCTS centres and at the Rescue Coordination Centres (RCC) and Marine Rescue Sub-centres (MRSC) to improve communication efficiency. To date, it is understood that the new digital immediate callback machines have been installed in all MCTS centres in the Newfoundland region. In addition, following the FLARE SAR operations, it was reported that SAR technicians (SAR TECHs) have now been provided with dry suits and full face masks. Furthermore, communication devices have been acquired to enable the SAR TECHs to communicate during future rescue missions. 4.1.4 Modification to the Canadian Coast Guard (CCG) Vessel W.G. GEORGE Following the occurrence, the W.G. GEORGE underwent modifications to address some problems encountered during her tasking. Rails were taken off a hatch located aft, providing more space to work on, and non-skid paint was applied to the decks. A Jason's cradle was fitted on the booms on each side of the vessel to aid in the recovery of persons in the water. A safety cable running on each side of the vessel has been fitted with dead eyes to allow the crew members on deck to better clip on their safety lines, thereby preventing them from being lost overboard in heavy seas. 4.1.5 Performance Standards for Thermal Protective Lifejackets - By the International Maritime Organization (IMO) In 1999, responding to the interests of several delegations, IMO's Maritime Safety Committee (MSC) approved recommendations on performance standards and tests for thermal protective lifejackets. Such lifejackets may be used in addition to, or in replacement of, SOLAS-approved lifejackets. When worn with warm clothing, such lifejackets are designed to provide thermal protection to limit the fall in body core temperature to 2C when the wearer is immersed for two hours in calm circulating water at a temperature of 10C. 4.2 Action Required 4.2.1 Stowage and Installation of Emergency Position Indicating Radio Beacon (EPIRB) The EPIRB reportedly carried on the FLARE either did not float free or did not self-activate as intended, and therefore failed to alert the SAR system of the distress. (No EPIRB signal was ever received from the FLARE, nor was the EPIRB recovered.) Although the MCTS at Stephenville, Newfoundland, received a MAYDAY distress alert, believed to have originated from the FLARE, the transmission was indistinct and incomplete. Consequently, the actual position of the sinking was not determined for some time. In severe climatic conditions, such as those encountered by the FLARE, it is essential that shore-based facilities be able to respond without delay. Time lost in the initial stages of an occurrence can be crucial to its eventual outcome. In 1988, a Canadian fishing vessel disappeared with no distress call having been received. It was more than five days before it was realized that the vessel was lost. The EPIRB carried on that fishing vessel was non-operative and was stowed in a locker. In contrast, in a January 1993 occurrence off Nova Scotia, the transmission from a float-free EPIRB was received moments after the vessel sank, and the Halifax RCC was able to launch a SAR operation within 10 minutes. The EPIRB carried aboard was an essential element in saving the lives of 11 of the 16 crew members. In most accidents involving bulk carriers since the early 1990s, the absence of any distress messages would indicate their loss was sudden and most likely due to structural failure, rapid flooding and loss of buoyancy/stability. As in this occurrence, most involved ships were at least 15 years old, and a high proportion were lost or had the potential to be lost through structural damage and/or heavy weather. Worldwide between 1990 and 1997, a total of 99 bulk carriers sank, with an associated loss of 654 lives. The IMO, recognizing the critical importance of timely alerting, identification and location of distressed vessels, requires the provision and availability of EPIRB signals. Chapter IV of the SOLAS, Consolidated Edition 1997, requires ships to carry EPIRBs which shall be installed in an easily accessible position and which shall be capable of floating free and automatically activating in emergency situations. In this occurrence, the investigation was unable to determine if the EPIRB was in its dedicated mounting on the starboard wing of the bridge. It is also unknown whether the unit floated free or not. The fact that no signal was received from the EPIRB contributed to the severity of consequences in this occurrence. Canadian SAR agencies expended substantial time and resources in attempting to save the lives of FLARE crew members. Valuable time was lost when SAR resources were misdirected to various assumed MAYDAY positions. Notwithstanding the substantial resources allocated to the FLARE SAR operation, it took six hours before the first survivors were spotted. (Airborne resources[7] searched an area of 4,371 square nautical miles, over 90 hours including the time spent in transit. Marine resources[8] searched an area of 1,702 square nautical miles, for close to 200 hours.) Had the FLARE managed to deploy her EPIRB and had it operated, it is likely that the early identification of an accurate position would have decreased the search time, thereby increasing the crew's chances of survival. Notwithstanding the SOLAS requirements on the carriage, stowage and proper installation of EPIRBs on vessels, risk to life continues to exist due to the unavailability of EPIRB signals. The Board is concerned that ship management personnel, ships' officers and crews may not be aware of the severe consequences of the improper stowage and installation of EPIRBs, thereby exposing themselves to undue risk in emergency situations. Furthermore, in view of the inherent weakness of relying on distress calls during an emergency and the loss of lives associated with delayed SAR operations, as evidenced in this occurrence, the Board recommends that: The Department of Transport, working through the appropriate agencies, advocate increased international measures aimed at ensuring that Emergency Position Indicating Radio Beacons are properly installed and deployable on vessels so that their distress signals are transmitted without delay in distress situations. M00-01 4.2.2 Immersion Suits for Operations in Cold Waters The North Atlantic Ocean is one of the most hostile environments in the world. Average mid-winter sea surface temperatures off the eastern seaboard range from 0C to 2C. The mid-summer temperatures range from 8C to 16C. In such harsh marine conditions, the survival time for a person immersed in water is often measured in minutes, while for a person wearing an immersion suit, survival time can run to several hours. People clad in such suits have been rescued following 18 hours of immersion in cold water. In February 1983, during a storm in the Atlantic Ocean, the United States bulk carrier MARINE ELECTRIC capsized and sank about 30 nautical miles east of Chincoteague, Virginia. Only 3 of the 34 persons on board survived. The United States National Transportation Safety Board (NTSB) investigated the occurrence and determined that the lack of personal thermal protection equipment for the crew--to minimize the effects of hypothermia--contributed to the heavy loss of life. Consequently, the NTSB recommended that exposure suits be provided for every person on board vessels that operate in waters where hypothermia can greatly reduce an individual's survival time[9]. The TSB has been concerned with the heavy loss of life associated with continuing losses of bulk carriers in and around Canadian waters. On 20 January 1990, the Cypriot bulk carrier CHARLIE, outbound from Montreal to Ceuta, Spain, was lost in the frigid Atlantic Ocean. No distress signal was received. The 27 persons on board went missing and were presumed lost. On 11 January 1991, while en route from Port-Cartier, Quebec, to Sweden, the Singaporean bulk carrier PROTEKTOR sank 225 nautical miles off St. John's, Newfoundland. All 33 persons on board were presumed lost. On 14 March 1993, the Liberian bulk carrier GOLD BOND CONVEYOR sank 110 nautical miles south of Nova Scotia in bad weather, with all 33 crew missing. On 01 January 1994, while outbound from Sept-les, Quebec, the Liberian ore carrier MARIKA sank 1,000 nautical miles east of Newfoundland. All 36 persons on board went missing and were presumed dead. Although all of these vessels were outbound from Canadian ports, because they were lost on the high seas, the TSB only assisted in the Flag State investigations. All of the above occurrences took place during the winter, in frigid North Atlantic waters. Under these conditions, without adequate thermal protection, the lives of the crew members would be at substantial risk due to hypothermia. Current SOLAS regulations do not require that an immersion suit be provided for each person on all cargo vessels. However, the regulations are such that an Administration may, at its discretion, require the provision of an immersion suit for each person on board. Canadian-flag ships, which regularly operate in higher latitudes, are required by regulations to provide at least one immersion suit for each crew member. International organizations, however, have not taken action to require that ships trading in colder climates provide an immersion suit for each crew member. In accordance with minimum SOLAS requirements, the FLARE was equipped with 6 immersion suits and 27 thermal protective aids; the latter items were stowed in the lifeboats. These thermal protective aids were intended for those persons, in open lifeboats, for whom immersion suits were not provided. All four surviving crew members of the FLARE, wearing lifejackets, were found to have been severely hypothermic and could barely move their limbs during their rescue, rendering the rescue operation difficult and subjecting SAR personnel to undue risks. Two other crew members, who had clung to the same lifeboat, remained alive for some time but succumbed to hypothermia before the survivors were spotted. (Neither immersion suits nor thermal protective aids were recovered during the SAR operation.) As described in the report, a sea survival model indicated that the use of immersion suits would have increased survival times to between 12 and 14 hours, depending on the clothing worn. The model also indicates that, in seawater of 2C, the best-clothed survivor would lose consciousness within 6.4 hours while the other survivors would reach this state in 2.0 to 2.3 hours. The Board believes that, under such conditions, crew survival largely depends on adequate thermal protection. Therefore, the Board recommends that: The Department of Transport advocate international measures requiring that an adequate immersion suit be provided for each person on board vessels operating in waters where hypothermia can greatly reduce an individual's survival time. M00-02 Furthermore, in rapidly developing distress situations such as those involving bulk carriers, it is critical that life-saving equipment, such as immersion suits, be readily accessible and rapidly retrievable without confusion. During this occurrence, the survivors of the FLARE indicated that they were unaware of where the suits were stowed, nor had they time to locate any of the immersion suits. In view of the frequency of occurrences involving bulk carriers that have rapidly capsized and sunk, often leaving crews insufficient time to avail themselves of on-board life-saving equipment, the Board further recommends that: The Department of Transport advocate international measures to help ensure that critical life-saving equipment, such as immersion suits and thermal protective aids, are stowed so that they are readily retrievable, without confusion, and that all crew members are familiar with their use and their stowage location. M00-03 4.2.3 Dynamic Loads on the Hull Due to Waves and Ship Motions The investigation found that the shallow forward draught, upon departure from Rotterdam as well as during the Atlantic crossing, made the vessel highly vulnerable to repeated pounding and slamming throughout the stormy voyage. Slamming, or the impact of the bow on the water during a large downward pitch, causes vibratory stresses or slamming stresses. For normal commercial vessels, the increase in stress due to slamming may reach 20 to 30 per cent of the primary sagging stress[10]. The most severe slamming and slamming/bending moments occur when the pitching period of a ship is approximately the same as the period of waves encountered. Slamming forces increase with increasing wave height and ship speed. The investigation concluded that the resulting severe whipping and flexing of the hull of the FLARE caused the sudden brittle fracture of the main deck and upper side shell plating. The light ballast loading condition of the FLARE, and the draughts recorded in Rotterdamprior to departure, confirm that the vessel sailed with a relatively shallow forward draught. The forward draught and the total quantity of water ballast on board were lower than those shown for the light ballast departure condition in the Loading Manual of the FLARE, which was provided for the guidance of the master. In addition to the light ballast loading condition, the vessel's Loading Manual also included a full or deep ballast condition suitable for longer and more exposed ocean passages such as this Atlantic crossing. While the vessel was making a winter crossing of the North Atlantic, the No. 4 hold/deep tank was not filled with ballast as indicated in the deep ballast loading condition of the vessel's Loading Manual. The total weight of ballast on board was significantly less than that stipulated in the vessel's Loading Manual. The manual indicated that all ballast tanks except the deep tank were to be filled to ensure forward and after draughts of 3.65 m and 7.0 m. The actual recorded forward draught was 0.58 m less than that shown in the Loading Manual for the light ballast departure condition. The International Convention on Load Lines, 1966 (LL1966) requires that the master of every ship be supplied with sufficient information to enable him to arrange for the loading and ballasting of the ship in such a way as to avoid any unacceptable stresses in the ship's structure. Such information is provided in detail in the Loading Manual. In addition, in 1998, the IMO adopted the Code of Practice for the Safe Loading and Unloading of Bulk Carriers (the Code). The Code was developed by IMO to minimise losses of bulk carriers due to structural failure resulting from excessive stresses. The Code requires that the ship be provided with a booklet which shall include , inter alia, ballasting and deballasting rates and capabilities, general loading and unloading instructions on the most adverse operating conditions during loading, unloading, ballasting operations and the voyage. Loading manuals provide masters with guidelines to assist them in ensuring that their vessels are safely ballasted and trimmed throughout a voyage, to maintain adequate structural integrity in various operating conditions. In addition to the vessel's Loading Manual, the Lloyd's Rules and Regulations for the Construction and Classification of Steel Ships that were applicable at the time the vessel was built make reference to the minimum forward draughts. According to these rules, the FLARE should have had a minimum forward draught of 4.6 m in order to avoid excessive forefoot exposure in rough seas. However, the vessel's actual forward draught of 3.35 m, reported to ECAREG three days before the occurrence, was substantially less. As indicated above, increasing the draught results in a decrease in both the slamming stress and the speed range in which slamming occurs. Had the instructions pertaining to ballasting and minimum draughts been followed, the vulnerability of the FLARE to pounding and slamming would have been markedly reduced, and the negative effects of dynamic stresses could have been avoided. However, the investigation was unable to determine why the instructions in the Loading Manual were not followed. The Board hopes that, with the effective implementation of and adherence to the International Safety Management Code (ISM), such deviations from norms may be minimized. In the interim, however, notwithstanding the existing guidelines and requirements of LL1966 and the IMO Code, it appears that structural failure of bulk carriers due to improper loading and distribution of ballast continues to occur. The Board is concerned that mariners may not fully appreciate that deviation from approved loading manuals may overstress the structure and lead to catastrophic failures. In particular, the Board is concerned that mariners may not fully appreciate the adverse consequences of dynamic loadings on the hull caused by slamming and bow flare impacts due to inadequate forward draughts. Therefore, the Board recommends that: The Department of Transport promote increased awareness and understanding in the international maritime community of potential structural failure associated with high frequency stresses on the hull due to slamming and pounding as a result of inadequate draughts of vessels operating in ballast conditions. M00-04 The Department of Transport, in coordination with international agencies (including the International Maritime Organization and the International Association of Classification Societies), bring the need for stricter adherence to approved loading manuals to the attention of shipowners, ship operators and ship masters in order to avoid undue structural stresses in bulk carriers. M00-05 The Transportation Safety Board will support and cooperate in the development of the submissions necessary to address these recommendations.